What is The Difference Between Host and Server?

The server is fundamentally different from the PC host. In information systems, servers are mainly used in databases and Web services, while PCs are mainly used in multimedia technologies and network terminals. The fundamental difference in application determines that the server has more powerful storage capabilities, faster failure recovery functions and faster recovery than PCs. Broad expansion space, and the server must also provide data backup functions that the PC host does not have. These technical indicators are essential elements of a server, and to a certain extent have become the standard for distinguishing between PCs and servers. These differences are also reflected in the four properties of the server ("Scalability" (scalability), "Usability" (availability), "Managbility" (manageability), "Availability" (practicability)) These four properties are usually abbreviated as "SUMA", which is a recognized standard for measuring server performance. The fundamental difference between a server and a PC host can also be reflected in these four characteristics.

Alibaba Cloud Simple Application Server:  Anti COVID-19 SME Enablement Program
$300 coupon package for all new SMEs and a $500 coupon for paying customers.

First of all, the server should be highly scalable. Scalable refers to the hardware configuration of the server, such as memory, adapters, hard disks, processors, etc., which can be easily added based on the original basis. The server has more PCI, PCI-X and other slots, more drive brackets and greater memory expansion capabilities. Provide redundant power supply and redundant fans, so that when the user's network is expanded, the server can also meet the new demand and protect the user's investment. In terms of server scalability, many brands now put forward the concept of "scaling on demand". Of course, this is mainly found in medium and high-end servers. For example, "XpandOnDemand" proposed in the IBM enterprise-level x-architecture server technology is such a technology. In middle and low-end servers, motherboards are usually required to provide multi-processor support. Although companies currently only need to use a single CPU, it is best to provide room for more than one CPU socket upgrade. For these small enterprises and emerging enterprises, relatively fast business growth is their outstanding feature, which puts forward special requirements for the scalability of the servers they use.

The availability of the server actually requires the server to have high reliability and high stability, not to crash or malfunction from time to time, and to minimize downtime for repair. The server is required to work continuously in most cases, so its reliability is very important, and it needs a quick way to restore the service in the event of a failure. When a PC has a system failure, methods such as restoring the registry and reinstalling the system are generally used to troubleshoot and restore the normal operation of the system. The loss of data is also limited to a single computer. The server is completely different. Many important data are stored on the server, and many network services are running on the server. Once the server fails, a large amount of data will be lost, and the damage caused is hard to estimate. Therefore, the requirements for server reliability are very high.

At present, a common method to improve reliability is redundant configuration of components and memory checking and error correction technology. For example, servers generally use ECC memory with checking and error correction capabilities, and some IBM servers also use special memory with ChipKill super checking and error correction capabilities, RAID technology, hot swap technology, redundant power supply, and redundancy. Fans and other methods make the server have fault tolerance and safety protection capabilities, thereby improving reliability. Of course, the most reminiscent "large chassis" is also necessary, because there are many devices in the server and generate more heat. Only a large chassis has enough space for heat dissipation to ensure the stability of the system.

In addition, the server also needs to have system and data backup functions, so it needs to provide at least the interfaces required by the data storage device. At present, tape is most commonly used for data backup, and the interfaces used by current tape drives are usually SCSI interfaces, so this requires a SCSI interface on the server motherboard to prepare for installation of tape drives (medium and high-end servers usually have Self-equipped tape drives) to realize effective backup of enterprise systems and data. It enables enterprises to use the most reliable one-key disaster recovery technology to recover the entire system in the shortest time when a server failure occurs, which provides a strong guarantee for the smooth operation of the server. Of course, the system and data backup can also use media such as optical disc library.

The practicality of the server is in terms of the processing capacity of the server. In most cases, the server requires high computing power and high processing efficiency. Therefore, the selected configuration is usually specially developed for the server, especially the CPU and motherboard. Choose server-specific. The server provides services for other computers on the network. Since the server needs to provide its data and hardware for network sharing, it needs to process a large amount of data when running certain applications, and it is likely to need to respond to dozens or hundreds of computers at the same time. Even thousands of computers request, so the computing speed of the server should be much faster than ordinary PCs, and the bandwidth should be much wider. At this time, the operating speed of the hard disk has higher requirements, so the server hard disk usually chooses the SCSI interface. The capacity is of course much larger than that of an ordinary PC.

At present, the most important technology is achieved through the use of symmetric multi-processing technology and cluster technology. Special technologies are usually used in the bus bandwidth technology, of course, this can be more reflected in medium and high-end servers.

The manageability of the server not only refers to the software, but also requires certain manageable performance in the hardware of the server, such as pre-alarm function and system monitoring function. The management performance of server hardware can be realized through ISC (Intel Server Control, Intel Server Control) technology. EMP (Emergency Management Port) technology is also a server management technology. This is also not available in PC mainframes.

Through the above comparison, it is not difficult to see that the server and the PC are strictly defined, and the function of the server is also impossible to replace by the PC. But then again, the server is actually a kind of computer after all. It was also developed by the PC mainframe. In the early days when the network was not very popular, there was no server name. At that time, there were only large computers and microcomputers in the entire computer field. There are two major categories of computers, but with the development and popularization of networks, especially local area networks, the middle-level computer "server" has begun to be accepted by the industry.